Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Akura]

Stupid question, but given it's a loss anyway I might as well ask. I think I found the problem with my keyboard. After noticing that it was specially the bottom rows of keys not responding plus the right side of the numpad and a few others, I traced the connections all the way back to the processor board. Noticed that the rubber keeping the pins in contact with the board is coming loose, especially on the outer edge. Sure enough, those are the pins leading to the non-functional keys.

As for the stupid question, is it safe to try using white glue or a hot glue gun to glue the rubber back into place? It doesn't need to hold for any more than a week. Of course I intend to avoid getting glue on the pins themselves.

[Starver]

For the processor speeds, I don't keep track any more, but different vendors of silicon used to take different approaches to how much processing capability was made available for various families of operation (esp. with lookahead calculations, what they'd specialise in pre-processing on the off chance it was still to be a valid request, so not thrown away), so different manufacturers (and, to greater or lesser extent, different product lines from the same manufacturer) could end up doing really well with sequential floating point vs array-wide integer ops, compared to one of the others.

DF famously does a lot of things differently to other consumer software (graphical rendering being to physics calcs in a  ratio quite unlike in FPSs, who often lean on GPUs' inate capabilities for the latter anyway, etc) so stress-tests by differing amounts the different elements.

edit: And the below reply too. In fact that's a better answer in all ways!


Keyboard-gluing... I've seen worse ideas, especially in the name of temporary fixing, but that's without being sure of your keyboard design. Sometimes there's a subtle layer of conductivity where it doesn't look like it, which you don't want to interfere with what was a conductive-adhesive join (though it reads like that isn't a case here) or spoil a capacitance effect of ome kind. Honestly, I'd be tempted to give it a go for myself (if I was in the scenario I think you are) but it could as easily react wrongly polymer-to-polymer for various reasons too. Trial a spot at one end first and monitor it for a duration of use?

[Rose]

Why does DF tend to run better on Intel processors compared to their AMD equivalents? I read this review of the Ryzen 3 3300X and 3100, and it shows that DF generates worlds faster on an i7-7700K (3 years older than the Ryzens). (Here's the link to that section of the review, if you're interested). To me, it's a bit strange, given that the Ryzens tend to be on par with (or in some cases, outperform) the Intel chip being compared (the aforementioned i7-7700K), at least in the review I read.

Why is this the case? Is it just down to single-threaded performance or having less cache, or is it more nuanced than that?

AMD tend to go for many weaker cores, while Intel go for fewer stronger cores.

DF mainly cares about single core performance, so by and large, it's better in Intel.

[Akura]

It appears gluing it worked, all keys tested appear to be working, with the exception of the 0 and . on the numpad. Not sure what's up there, but it's still a far better improvement than before.

[Lord Shonus]

I don't know what causes it, but Windows keeps lowering the specific application volume on my web browser and Discord to nothing. It may have something to do with unplugging headphones, but even that isn't desired behavior. Anyone know a way tos top this?

[wierd]

Afraid not.  For me, the annoyance is microsoft repeatedly "UPDATING!!" my wifi drivers to one that causes network dropoffs and bluescreens. 

This is the third time I have directly installed the Intel driver pack for that chipset.  I *REALLY* (demonic, gravel tone) want them to *STOP* fucking with my drivers every time they want to install updates.

Look, I fucking get it that some people don't know what they are doing and will use a driver from 1992, ok?  However, if I installed a driver package from Intel, it means I want to use the drivers that have gone through Intel's QC process, since microsoft no longer has a QC department. Got it, Microsoft?  I DO NOT LIKE YOUR UNTESTED BULLSHIT, STOP CHANGING IT.  I DO NOT CARE THAT THE VERSION NUMBER IS HIGHER.

(rage.)

Anyhow..  I dont think there is a way to forcibly prevent window from "helpfully" reducing the volume to zero for you. 

[Reelya]

I don't know what causes it, but Windows keeps lowering the specific application volume on my web browser and Discord to nothing. It may have something to do with unplugging headphones, but even that isn't desired behavior. Anyone know a way tos top this?

Use the universal solvent for Windows problems. AutoHotKey.

https://www.autohotkey.com/boards/viewtopic.php?t=46654

So what you could do with AutoHotKey is run a time interval loop and have that reset the volume on those applications to 100% each time. Say every 5 seconds or something.

[Starver]

Having trouble with the bay12games.com domain.  But it's not a Bay12 issue so far as I can tell, so putting it here to get ideas...

Compare and contrast:
1) Old Laptop (Wifi connection to home ISP router): bay12games.com (with www. or dffd. subdomains) redirected to my ISP's default Error Replacement Service (looks like an unbranded Google results for "what they they I want to connect to", mostly Paid Ads but with bay12games, bay12forums and the Patreon/bay12 pages in there for good measure.  Confirmed through the likes of Ping and Tracert that traffic goes to this service.  Even after I 'activated' the ISP's opt-out, which quite famously does not opt out.

2) Old PC (wired connection to home ISP router): no ad-jacking DNS, but this may be because I changed the default DNS used to something that helped a different issue. However, the browser (Firefox) nw gives the old "did you mistype/miss out the www. part?" message and a Try Again button that does nothing more, so obviously somewhere in the upstream DNS I still get something invalid.

3) Tablet, with Mobile Internet or Wifi to the same router: bay12games.com works as it always has, on either mobile (different ISP, not even one under the same umbrella due to merger/corporate assocation) or wifi.  Not sure if this is because it got the working DNS info while on mobile ISP and continues to use it successfully over the home ISP, because I don't have easy access to any of the same tools for which I discovered such inner details of the Error Replacement as I have already found.


No problems with any other domains (bay12forums included) but the two offshoots bay12games I tried. I would guess (without easy tracking other than resorting to third party tools) that dffd is not even hosted on the same hardware, so why both would (initially) even return an NXDOMAIN, for it to then selectively replace with the 'helpful' URL hijack on just the one machine that does that I have no idea.

The ISP involved is notoriously bad on support (via all lines of communication). I could pester them by phone, email, livetalk, at least two different support/community forums, etc, but noting recent traffic discussing the inability to opt-out with the opt-out (and unrelated VPN troubles that this is causing, which other are having) I'm wondering what things I might have overlooked that I can test and/or tweak.


(I could perhaps (re)change the DNS being called upon, or something more drastic, but I don't want to do that lightly. Not for just bay12games access, which I don't specifically need right now and I only discovered when I got rerouted after having not typed the /smf/ ending to the bat12forums URL, in haste. I can review the rest of the site on the tablet, and have been, and there's nothing I currently need to download to Windows.)

If I was free to move equipment to a different place, I might take the laptop over to try to get through a different ISP (from the current location or mobile access) and do a suitable /refresh on the DNS info but, right without the ability to do this anywhere but its current locale, it seems it would just refail the same apart from any redirect element. Can't work out how to possibly 'break' the tablet in the same way (if it is only true right now because it recalls a working server resolution, not yet deemed needing auto /refresh) and don't much care to.


Do I make sense? Maybe not perfectly, but it's a puzzle, in case anyone wants to help me solve it before I get around to doing anything too drastic myself, or it resolves itself spontaneously.

[Ziusudra]

Actually it is/was a bay12 issue. https://www.reddit.com/r/dwarffortress/comments/gtc692/the_bay12games_website_seems_to_be_down/

[Starver]

Well, it was available by the new tablet. Didn't drill too deep (the front screen and the /Dwarves/ inner-home page). Typically, just dug up an older tablet (younger than the Windows hardware, only Wifi) and got the ISP's 'helpful' page, thought it was a new data-point, but then refreshed the page on the new tablet (on Wifi) and ISP intervenes.

Switched back to mobile ISP, top-home Ok (Dwarf Fortress, Liberal Crime Squad, WWI Medic, etc) and then gone to pages I've probably never ever visited on New Tablet like WW1Medic screenshot. Either it's very aggressive prefetched-and-optimised-for-mobile practice by mobile ISP or I've got a fully working/fixed DNS by that provider.

Mobile off, Wifi on, look at other WW1Medic screenshot. Yep, still there (i.e. totally new info from probably-remembered looked-up end-point, and not sourced from an overly-smart intermediary proxy or anything - unless it was pre-fetched by the tablet itself while on Mobile, but that looks is inconsistent with its also rather annoying data-saving features that I'm sure I haven't managed to completely turn off).

So, I'm positive it's absolutely not the server itself, if anyone can make use of this assumption, but that something within the Intertubes is causing selective NXDOMAINing (or else an anti-problem is getting around the legitimate error) and the landline ISP's opt-out issue is just arising from that and 'only' a completely different issue.


Been ages since I ever messed about seriously all with that backbone stuff (and this was with MX records, mostly) so I'm probably going to leave it at that and leave the redditers(?) to deal with it.  Well, there actually are far worse options.

Anyway, apologies for spilling so much extra ASCII over this issue. It sure is nice to have a really meaty tech problem of my own.


edit: Five minutes later, still on Wifi, tablet stops remembering the right destination (DNS data rechecked, found wanting, I presume), back to mobile and I can access Kobold Quest via the Other Games springboard of the home screen - a definitely rarely-visited page.  Data point.  Last one.

[methylatedspirit]

I'm building a PC, and I can't get it to turn on. I got the RGB fans to spin one time, but after unplugging and replugging everything, it doesn't work at all.

What could I have done wrong here? It could be as simple as a header plugged in to the wrong place, since I legitimately have no idea what's going on.

[wierd]

Unplug the cord from the back of the PSU. Flip the switch on the back. Press in the power button on the PC (with it still unplugged).  Wait 5 seconds.  Flip the switch back. Plug it back in.

Re-attempt power-on.

Rationale:

There is protection circuitry inside the PSUs of modern systems. This circuitry prevents energizing the main power lines if there is spurious voltages on any of the connections, and the like. This can happen from any number of causes, but the biggest one is a power spike, either from ESD or from the power cord being yanked, a rocker switch sending back a sharp voltage curve, etc.

Flipping the switch on the back of the PSU disengages certain circuits in the PSU (and basically turn it off).  Holding in the power button forces whatever voltage is retained in the large filtering capacitors in the PSU to fully discharge.

Returning the switch back to the operational setting, then reattempting power on after this process "resets" this protection circuitry.

[methylatedspirit]

I don't seem to be having any luck getting it to work. There was a moment where the RGB fans lit up, but they quickly turned back off.

[wierd]

Do you have a different PSU? If so, try that.  If not, begin removing all PCIe and PCI devices from the system, and unplug all power and data cords going to disk drives, then try again.

[methylatedspirit]

I got some help from my dad, and he managed to get it to work, somehow. Maybe a loose power cable somewhere, I dunno. At least it's working now, running Windows 10.